Occupational and Environmental Medicine 1995;52:611-620 61

Biological monitoring of exposure to benzene: Occup Environ Med: first published as 10.1136/oem.52.9.611 on 1 September 1995. Downloaded from a comparison between S-phenylmercapturic acid, trans,trans-muconic acid, and phenol

P J Boogaard, N J van Sittert

Abstract life S-PMA is also a more reliable bio- Objectives-Comparison of the suitabil- marker than tt-MA for benzene expo- ity of two minor urinary metabolites sures during 12 hour shifts. For biological of benzene, trans,trans-muconic acid monitoring of exposure to benzene con- (tt-MA) and S-phenylmercapturic acid centrations higher than 1 ppm (8 h TWA) (S-PMA), as biomarkers for low levels of tt-MA is also suitable and may even be benzene exposure. preferred due to its greater ease of mea- Methods-The sensitivity of analytical surement. methods ofmeasuring tt-MA and S-PMA were improved and applied to 434 urine (Occup Environ Med 1995;52:611-620) samples collected from 188 workers in 12 studies in different petrochemical indus- tries and from 52 control workers with no Keywords: benzene; biomarkers; S-phenylmercapturic occupational exposure to benzene. In acid; transtrans-muconic acid nine studies airborne benzene concentra- tions were assessed by personal air moni- The main metabolites that can be found in toring. urine after exposure to benzene are phenol Results-Strong correlations were found and its conjugates.' As a consequence, total between tt-MA and S-PMA concentra- urinary phenol has been used as a biomarker tions in samples from the end of the shift for occupational exposure to benzene for a and between either of these variables and long time. Its use, however, is limited to ben- airborne benzene concentrations. It was zene concentrations over 5 ppm (16 mg/M3) calculated that exposure to 1 ppm (8 hour as an 8 hour time weighted average (TWA) time weighted average (TWA)) benzene because a substantial background of phenol leads to an average concentration of 1'7 may be found in non-exposed people due to mg tt-MA and 47 jg S-PMA/g creatinine dietary habits. In many countries, however, in samples from the end of the shift. It the occupational limit concentration for air- was estimated that, on average, 3 9%/6 borne benzene will be, or has already been,

(range 1.90/o-7.3%) of an inhaled dose of lowered to 1 ppm (8 h TWA). In some coun- http://oem.bmj.com/ Shell Research BV, benzene was excreted as tt-MA with tries even lower values have been adopted Shell Molecular Toxicology, an apparent elimination half life of (Sweden 0 5 ppm) or proposed (United Badhuisweg 3, 5*0 (SD 2.3) hours and 0*11% (range States 0 3 ppm). As a consequence new bio- PO Box 38000, 0-050/%0-26%) as S-PMA with a half life markers are needed. Although measurement 1030 BN Amsterdam, The Netherlands of 9 1 (SD 3.7) hours. The mean urinary of benzene itself either in blood or exhaled air P J Boogaard S-PMA in 14 moderate smokers and is both highly specific and sensitive,23 urinary N J van Sittert 38 non-smokers was 3*61 and 1*99 uglg measurements are preferred in occupational on September 27, 2021 by guest. Protected copyright. Shell Internationale creatinine, respectively and the mean medicine as collection of urine is not only Petroleum was 0 058 and 0 037 mglg non-invasive but also easily performed even at Maatschappij BV, urinary tt-MA Health, Safety, and creatinine, respectively. S-PMA proved distant working locations. Over the past few Environment to be more specific and more sensitive years two minor urinary metabolites of ben- Division, Biomedical = 0-030, Fisher's exact test) than tt- zene have been introduced as new biomarkers Services, PO Box 162, (P 2501 AN The Hague, MA. S-PMA, but not tt-MA, was always for determination of benzene exposure: The Neterlands; and detectable in the urine of smokers who trans,trans-muconic acid (tt-MA, 2,4-hexadi- Shell Biomedical were not occupationally exposed. S-PMA enedioic acid)4 1 and S-phenylmercapturic Laboratory, Shell Nederland was also detectable in 20 of the 38 non- acid (S-PMA, N-acetyl-S-phenyl-L-cys- Raffinaderij/Chemie smokers from the control group whereas teine)." 14 We applied these two markers in a BV, Occupational tt-MA was detectable in only nine ofthese series of studies in industrial workers with Health and Hygiene to benzene to determine Division, PO Box samples. The inferior specificity of tt-MA potential exposure 3000, 3190 GA is due to relatively high background val- their comparative suitability for biological Hoogvliet ues (up to 0'71 mglg creatinine in this monitoring. (Rotterdam), The in Netherlands study) that may be found non-occupa- exposed people. Correspondence to: tionally Dr P J Boogaard, Conclusions-Although both tt-MA and Methods and materials Shell Research BV, are sensitive only POPULATIONS AND COLLECTIONS OF URINARY Shell Molecular Toxicology, S-PMA biomarkers, Badhuisweg 3, PO Box S-PMA allows reliable determination of AND AIRBORNE SAMPLES 38000, 1030 BN down to 0 3 ppm (8 h For the estimation of normal values of urinary Amsterdam, benzene exposures The Netherlands TWA) due to its superior specificity. tt-MA, a study was conducted in 46 employ- Accepted 15 May 1995 Because it has a longer elimination half ees from the technical and six employees from 612 Boogaard, Sittert

Summary ofstudies to monitor benzene exposure Benzene in air Urinary nA-MA (begin) Occup Environ Med: first published as 10.1136/oem.52.9.611 on 1 September 1995. Downloaded from Number of (mg/Mr3, 8 h TWA) (mg/g creannine) Working measurements Study Activity Country day (workers) Median (Range) Median (Range) 1 Maintenance of North Sea 1 4 (4) 8-4* (<0-1-19-2) 0-06 (0-01-0-23) gas production 2 4 (4) 1-8* (<0-1-3-6) 0-68 (<0-01-3-26) platforms all 8 (4) 1-8* (<0-1-19-2) 0-06 (<0-01-3-26) 2 Maintenance of North Sea 1 3 (3) < 0-1* (<0-13) 0-33 (0-16-0-61) gas production 2 3 (3) 10-1* (7-5-10-7) 0-40 (0 26-0 65) platforms all 6 (4) 3-8* (<0-1-10-7) 0-37 (0-16-0-65) 3 Activities before North Sea 1 11 (11) 0-98* (<0-1-17-9) 0-16 (0-03-1-03) shut down on 2 11 (11) 0-65* (<0-1-15-6) 0-24 (<0-01-1-48) gas production 3 11 (11) 0.33* (<01-3 6) 0 40 (<001-1 88) platform 4 9 (9) 0.10 (0-01-0-24) all 42 (12) 0-65* (<0-1-17-9) 0-17 (<0-01-1-88) 4 Routine operation North Sea 1 4 (4) 0-65* (0 3-2 9) 0-30 (0-06-0 67) on gas production 2 4 (4) - 0-35 (0-06-1-36) platform 3 4 (4) 2-60* (0.7-6 5) 0-20 (0 06-0 27) 4 4 (4) < 0-1* (< 0-1-1-0) 0-56 (0-30-0-71) all 16 (4) 0-85* (<0-1-6-5) 0-24 (0-06-1-36) 5 Chemical Netherlands 1-5 146 (69) 1-26 (<0-01-211-1) 0-15 (<0-01-0-80) manufacturing 6 Chemical Netherlands 1-5 41 (22) 17t (<0-01-201-5)t - manaufacturing 7 Chemical Germany 1-5 39 (39) - manufacturing 8 Chemical France 1 2 (2) - (0-33-2 28) - (0 05-0 86) manufacturing 2 2 (2) - (< 0-01-1-95) - (0 08-0-09) 3 2 (2) - (0 33-5 20) - (0-01-0-14) 4 2 (2) - (0-98-3 25) - (0-01-0-45) 5 2 (2) - (< 0-1-0-98) - (0 20-0 20) 6 2 (2) - (1-63-1-63) - (0-01-0-07) 7 2 (2) - (<01-< 0-1) - (0 52-0 34) 8 2 (2) - (<0-1-< 0-1) - (0-01-0-08) all 16 (2) 0-65* (<0-1-5-20) 0-08 (0-01-0-86) 9 Oil refinery with France 5 8 (8) aromatics plant 10 Oil refinery with Germany 1-5 14 (8) 0 77 (0-11-3-30) aromatics plant 11 Road tanker Belgium 1 14 (14) - - 0-08 (0-01-0-40) loading 2 14 (14) - - 0-06 (0-01-0-23) (top filling) 3 14 (14) - - 0-10 (0-01-0-42) all 42 (14) - - 0-06 (0-01-0-42) 12 Pipe line Netherlands 1 2 (2) - (1-3-1-5) - (0-04-0-11) maintenance 2 2 (2) - (0-8-1-4) - (0-13-0-14) 3 2 (2) - - (0-10-0-22) all 6 (2) - (0-8-1-5) - (0-04-0 22) *12 Hour shifts; personal air monitoring was not performed on all operators.

the medical department who were not occu- (seven samples) who had airborne exposure to pationally exposed to benzene (controls). This benzene higher than 1 5 ppm (12 h TWA) control group comprised 14 smokers and 38 urinary phenol concentrations were also http://oem.bmj.com/ non-smokers. measured. For the comparison of the methods, 12 Study 3 was performed on 16 people studies in eight locations (four countries) have engaged in activities before shut down. Only been performed in the period from 1992 to five people were involved for one day, the 1994 on workers who were potentially other 11 people were involved for two to four exposed to benzene during manufacturing consecutive 12 hour shifts. Personal air sam- and maintenance in natural

operations gas pling was carried out on these 11 people as on September 27, 2021 by guest. Protected copyright. production installations, in refineries, and in described for study 1. All 16 people provided chemical plants (table). urine samples before and after the shift on Four surveys were carried out in 1992 dur- each working day and nine of the 11 people ing various maintenance operations on natural who worked more than one day provided an gas production platforms in the North Sea. additional urine sample 12 hours after the end Studies 1 and 2 were both performed on of the last shift. One worker did not provide a two consecutive days, each on four workers sample at the end of the shift on day 3. Both involved in small maintenance operations. S-PMA and u-MA were determined in all 83 Personal air monitoring was carried out on urine samples. Also phenol was measured in these eight workers for the whole 12 hour shift 18 urine samples of workers with airborne with charcoal tubes and personal sampler benzene concentrations higher than 2 mg/M3 pumps (flow rate about 100 mi/min). The (12 h TWA). detection limit of airborne benzene was 0-1 Study 4 was performed on four consecutive mg/M3 (12 h TWA). The occupational days (12 hour shifts) on four people involved hygienist involved noticed skin contact in in routine small maintenance operations. study 1 on four occasions (three people). All Personal air monitoring was carried out as the participants provided urine samples before described for study 1. From all workers urine and after the shift for determination of S- samples were collected before and after the PMA and tt-MA (in total 16 samples for study shift for measurement of S-PMA and tt-MA 1 and 12 samples for study 2). In the urinary (32 samples in total). For two people with samples of two operators from study 1 (eight relatively high benzene exposures phenol samples) and three operators from study 2 measurements were also made. Biological monitoring ofexposure to benzene: a comparison between S-phenylmercapturic acid, trans,trans-muconic acid, and phenol 613

handled benzene. Exposure to benzene was Occup Environ Med: first published as 10.1136/oem.52.9.611 on 1 September 1995. Downloaded from Uninary ti-MA (end) Urinary S-PMA (begin) Urinary S-PMA (end) monitored by measurement of urinary tt-MA (mg/g creatinine) (liglg creatinine) (pg/g creatinine) and S-PMA. From all workers a single urine Median (Range) Median (Range) Median (Range) sample was collected at the end of the 8 hour shift. No air was 4-62 (< 0-01-12-5) < 3 (< 1-5) 151 (< 1-798) personal sampling per- 1-05 (< 0-01-1-30) 145 (2-496) 35 (3-70) formed. 1-05 (<0-01-12-5) 5 (< 1-496) 35 (< 1-798) 8 involved two in a chemical 0-48 (0-13-0-72) 12 (4-104) 15 (3-23) Study people 2-59 (2-48-6-23) 20 (14-34) 131 (64-175) plant in France who were biologically moni- 1-60 (0-13-6-23) 17 (4-104) 44 (3-175) tored for benzene on consecu- 0-67 (< 0-01-4-06) 3 (< 1-23) 18 (5-392) exposure eight 0-77 (0-14-2-63) 17 (4-143) 15 (1-60) tive days by measurement of tt-MA and 0-37 (0-07-1-28) 19 (5-43) 12 (3-51) S-PMA in urine samples from before and - - 7 (2-17) - 0-40 (< 0-01-4-06) 7 (1-143) 12 (1-392) after the shift. In 12 selected samples phenol 0-35 (0-14-0-76) 3 (2-17) 6 (5-25) was also measured. Personal air was 0-22 (< 0-01-0-28) 6 (5-49) 4 (< 2-10) sampling 1-07 (0-24-1-70) 2 (2-9) 11 (5-66) carried out in all 12 hour shifts, as in study 5. 0-43 (< 0-01-0-85) 6 (5-45) 10 (1-15) Two studies were carried out in 1992 in oil 0-24 (<0-01-1-70) 4 (2-49) 6 (< 2-66) 1-53 (<0-01-20-78) 7 (2-46) 36 (2-2312) refineries with aromatic plants in France and Germany. Study 9 was carried out on eight - 38 (20-1175) 1-61 (0-13-39-49) people who collected a single urine sample at 1-08 (0-18-18-18) - 12 (3-277) the end of the shift. In the samples tt-MA, - (0-26-0-09) - (2-3) - (11-22) S-PMA, and phenol were measured. No air - (0-11-1-97) - (1-12) - (4-66) measurements were carried out. Study 10 was - (0-01-4-20) - (2-40) - (4-133) - (0-08-2-42) - (1-92) - (3-143) performed on eight workers who provided - (0-15-0-81) - (1-31) - (1-41) urine samples at the end of the shift once, on - (0-74-0-35) - (4-20) - (18-17) - (0-01-0-01) - (21-22) - (9-5) the last day of the week (two people), or - (0-01-0 05) - (1-5) - (< 2-12) on the third and fifth of the five 0-13 (0-01-4-20) 5 (1-92) 12 (1-143) twice, day day 0-18 (0-01-0-48) - 26 (3-85) working week (six people). Both tt-MA and S-PMA were measured in all - 12 (2-98) samples, phenol 0-55 (0-11-1-51) was measured in three selected samples. Air 0-17 (0-01-0-53) 1 (< 1-11) 6 (1-13) was pumps 0-28 (0-08-0-61) 4 (1-14) 5 (1-34) sampling by personal sampler 0-20 (0-05-1-05) 5 (1-13) 5 (2-38) (flow rate between 40 and 80 ml/min) with 0-21 (0-01-1-05) 3 (< 1-14) 6 (1-38) charcoal as the absorbent. The detection limit - (0-50-0-57) - (2-4) (3-21) (0-56-0-57) (2-14) (3-24) was 0-01 mg/M3 (8 h TWA). - - - (1-21) - 11 was in 1992 in (3-24) Study performed (0-50-0-57) (1-21) Belgium on 14 drivers of road tankers who were involved in loading the tanker (top fill- ing) with petrols containing an average of 3% v/v of benzene."5 The drivers were biologically monitored by measurement of tt-MA and Four studies were carried out during rou- S-PMA in urine samples collected at the tine operations in chemical plants. beginning and the end of the 8 hour shift on Study 5 took place in 1992 in The three consecutive days. Air monitoring of ben- http://oem.bmj.com/ Netherlands and involved 69 people who zene was performed as in study 5. Individual worked in the production of ethyl benzene data were not available. from benzene. Forty of them were biologically Study 12 was carried out in The monitored only once by measurement of uri- Netherlands in 1994 on two men involved in nary metabolites in samples collected both at cleaning and maintenance operations for two the beginning and the end of the eight hour days on a pipeline system used to transport

shift (seven people) or only at the end of the naphthas with a high content of aromatic on September 27, 2021 by guest. Protected copyright. shift (33 people). The other 29 people were compounds. The men were biologically moni- biologically monitored on between two to 12 tored by measurement of tt-MA and S-PMA different days. In total 169 samples were col- in urine samples taken before and after the lected and analysed for S-PMA and tt-MA, in shift and at the beginning of the third day. Air 115 samples phenol was also measured. sampling during the activities was carried out Personal air sampling was performed in 44 with charcoal tubes and personal sampler workers for the full shift on one or more days pumps (flow rate about 200 ml/min). (90 samples in total) with passive dosimeters In all studies, urine samples were collected (3M gas diffusion batches, type 3500). The in polythene bottles and acidified to pH 2 detection limit of airborne benzene was 0 1 with 6 M hydrochloric acid on site. Stability mg/M3 (8 h TWA). studies of tt-MA, S-PMA, and phenol in urine Study 6 was carried out in the same plant in had shown that under these conditions con- 1993 on 22 people who provided 41 urine centrations did not change if stored at room samples collected at the end of the 8 hour temperature or 40C for at least a month.9 13 shift on between one and four different days. The samples were transported to the Shell In all samples tt-MA and S-PMA were deter- Biomedical Laboratory, Rotterdam, for mea- mined, but no measurements of phenol were surement of urinary benzene metabolites and made. Personal air monitoring was carried out creatinine. Urinary creatinine was measured as in study 5 on eight people (13 samples in to adjust for the variations in concentration of total). urine samples due to differences in fluid Study 7 was carried out in 1993 on 39 intake between workers. Urinary creatinine workers in a chemical plant in Germany who was previously shown to provide a good 614 Boogaard, Sitert

correction for spot samples if 24 hour samples coefficient of variation of replicate analyses could not be collected'3 and it concomitantly (n = 10) was 3 2% at a spiked concentration Occup Environ Med: first published as 10.1136/oem.52.9.611 on 1 September 1995. Downloaded from allows adjustment for the variable dilution of 0-64 mg/l and 0-4% at a spiked concentra- introduced by the acidification of the samples. tion of 5-55 mg/l. The reproducibility was Personal air sampling and benzene analyses tested by analysis of a set of identical samples were performed by local staff. in a series (n = 6) of runs. The coefficient of variation was 11% for samples spiked with MEASUREMENT OF URINARY METABOLITES 1-1 1 mg n-MA/l and 2-7% for samples spiked The measurement of phenol and creatinine with 5-57 mg u-MA/l. With this procedure, have been described previously.'3 The up to 40 urine samples could be analysed in S-PMA was also measured as described previ- one day. ously,'3 but as an internal standard deuterium labelled S-PMA was used instead of MEASUREMENT OF BENZENE IN AIR S-benzylmercapturic acid (S-BMA). The S- The benzene vapours adsorbed on the gas dif- [pentadeuterophenyl]mercapturic acid (d5-S- fusion badges or the charcoal tubes during PMA) was synthesised from ring-d5 aniline personal air sampling were desorbed by car- with the Gattermann reaction according to bon disulphide and the solution was analysed RumpPf6 and was recrystallised twice from with a gas chromatographic method with ethanol with water. The use of d5-S-PMA as flame ionisation detection according to stan- an internal standard allowed a limit of detec- dard methods published by the National tion of 1 jg/l. A series of urine samples were Institute for Occupational Safety and Health pooled, spiked with S-PMA at two different (NIOSH).'7 concentrations, acidified, and stored frozen at - 20'C. In all runs a set of samples from this CALCULATIONS AND STATISTICS urine pool was analysed to check the repro- The cumulative excretions of tt-MA and S- ducibility of the method. The coefficient of PMA were estimated from the area under the variation between different runs (n = 42) was curve of urinary excretion v time plots and 8-9% for the pool with 28 pug S-PMA/l and were used to estimate the percentages of the 6d1% for the pool with 104 pug S-PMA/l. The respiratory benzene dose excreted as tt-MA coefficient of variation of replicate analyses and S-PMA in the urine. For 14 workers, (n = 10) was 2 5% and 3 5% at spiked con- mainly from studies 3, 4, and 8 personal air centrations of 13-1 and 50 1 pg S-PMA/l, monitoring data as well as sufficient biological respectively. With this procedure up to 30 monitoring data were available both to esti- urine samples could be analysed in one day. mate the respiratory benzene doses and to cal- The measurement of tt-MA was essentially culate the areas under the curve. For the that of Ducos et al.5 In our hands, this method calculation of these areas the linear trape- gave non-reproducible recoveries. It seemed zoidal rule was used in the absorption phases that the recovery was dependent on both the and the logarithmic trapezoidal rule in the urinary pH and the brand of the quaternary elimination phases. The following formula ammonium ion exchange resin (SAX was used for the logarithmic trapezoidal rule:

columns) used. With some brands poor recov- http://oem.bmj.com/ eries were obtained with slightly acidic urines. area under the curve(1 2) = 0.0625 x Therefore, the urine samples were brought to ((c - C2) x t2 - t,))/(ln(c,) - ln(C2)), pH 7-10 by addition of 35% (w/v) aqueous where 0-0625 is the average urinary volume sodium hydroxide solution before the sample (1/h), c, denotes the concentration of metabo- was cleaned up by solid phase extraction. An lite (either tt-MA or S-PMA) in the urine aliquot of 1I00 ml of urine was subsequently sample collected at time t, (end of the shift) passed through a SAX column. The column and the concentration c2 of metabolite at time on September 27, 2021 by guest. Protected copyright. was washed with 3 ml 1 % (v/v) aqueous acetic t2 (next morning) during the elimination acid and the tt-MA eluted with 4 ml 10% phase (c, > c2). The absorbed benzene dose (v/v) aqueous acetic acid into calibrated tubes was estimated with the formula: and the volume was made up to 5 0 ml with purified water. Aliquots (20 pl) of this solu- dose = 0-47 x C x V, tion were analysed by high performance liquid where 0 47 is the fraction of benzene that is chromatography (HPLC) with UV (A = 259 retained upon inhalation,'8 C denotes the air- nm) detection (Hewlett Packard 1084B) at borne concentration (mg benzene/m3) and V 20°C with a 100 x 2-1 mm stainless steel col- the average respiratory volume (1 25 m3/h). umn with Spherisorb 5 ODS-2 as the immo- Both for tt-MA and S-PMA apparent uri- bile phase and 20% (v/v) methanol in 1% nary elimination rate constants were calcu- (v/v) aqueous acetic acid as the mobile phase. lated from workers who had provided urine The flow rate was set at 1 ml/min. For cali- samples at the beginning and end of a shift on bration, a series of tt-MA standards in fresh at least two consecutive days. For the calcula- control urine samples were analysed together tion of the apparent elimination rate constants with the urine samples. Alkalisation of the (k-) it was assumed that the absorption was urine samples before the solid phase extrac- completed and that first order elimination tion greatly improved the reproducibility: kinetics were obeyed during the collection optimum recoveries (invariably over 95%) period. The following formula was used: were obtained if the urinary pH ranged from 7 to 10. A more effective clean up was a side ln(c2/c,) = - ke x (t2 - t.), effect of the alkalisation of the samples, which where c, denotes the concentration of meta- lowered the detection limit to 0-01 mg/l. The bolite (either tt-MA or S-PMA) in the urine Biological monitoring ofexposure to benzene: a comparison between S-phenylmercapturic acid, trans,trans-muconic acid, and phenol 615

sample collected at the end of the shift at time Results t, and c2 the concentration of metabolite at URINARY tt-MA AND S-PMA IN WORKERS WITH Occup Environ Med: first published as 10.1136/oem.52.9.611 on 1 September 1995. Downloaded from the beginning of the next day shift at time t2. NO OCCUPATIONAL EXPOSURE Occasionally, when the absorption phase was The 38 non-smokers in the control group had not yet completed at the end of the shift (t,), a mean (SEM) urinary tt-MA concentration c, was higher than cl. In these cases k, was cal- of 0 037 (0-016) and the 14 smokers 0-058 culated from the decline in metabolite during (0-013) mg/g creatinine. The mean (SEM) S- the same shift but only if there was no PMA concentration in the 38 non-smokers detectable exposure to benzene. In these cases was 1 99 (0 29) and in the 14 smokers 3X61 c, denotes the concentration of metabolite (0 57) ,ug/g creatinine. S-PMA was present at (either tt-MA or S-PMA) in the urine sample detectable concentrations in the urine of all collected at the beginning of the shift (at time smokers and in 20 of the 38 non-smokers. tQ) and c2 the concentration of metabolite at This showed the superior sensitivity (P = the end of the same shift (at time t,). For most 0-030, Fisher's exact test) of S-PMA over workers values of k, could be calculated more tt-MA, for which urinary concentrations were than once; in these cases the separate values below the detection limit in three smokers and were averaged. From the averaged values the 29 non-smokers. Although the delectability corresponding half life (t,12) was calculated for (yes, no) of both biomarkers was strongly cor- each worker with the following formula: related with smoking (yes, no) (P < 0000 1, Fisher's exact test), multivariate regression t112= - ln(2)/k,. analysis disclosed no correlation with the For all statistical calculations the mainframe number of cigarettes smoked and either S- version of the SAS software package (version PMA or tt-MA. 6-07 or 6&08) was used. For S-PMA and tt- MA results below the calculated value for the URINARY tt-MA AND S-PMA IN limit of detection of the respective methods OCCUPATIONALLY EXPOSED POPULATIONS half of this value was substituted. The table gives an overview of the results obtained in the 12 studies in industrial workers with potential exposure to benzene. The high 150 - airborne benzene concentrations measured in -A 140 studies 1, 2, and 3 (up to 19-2 mg/m3) did not 130 o ii-MA _5 C-E reflect normal operations but peak exposures - 120 t * S-PMA CD measured in selected operators with relatively aZ 110 8 hTWAbenzene a) high potential exposure to aromatic com- C a) 4 N pounds during special maintenance opera- C 100 c ._- -4 a) tions on gas production platforms. Both 90 -4 tt-MA and S-PMA excretion increased over ' 80+- ._ cm -3 C the shift and reflected the airborne benzene C routine operations (study 4) 70L cu exposure. During < 60- benzene exposure was usually far below 2 -2 1 ppm (3-25 mg/m3) but, nevertheless, http://oem.bmj.com/ 0 50t- 40k~- increases in both tt-MA and S-PMA were -E found. In the urine samples provided by oper- 1 ators with relatively high benzene exposures 0 00 (> 0-5 ppm) phenol was also measured. It that correlated well with both o 0 seemed phenol 20 8 20 o -c tt-MA and S-PMA in operators with low

Time (h) E background values of phenol (< 5 mg/g creati- on September 27, 2021 by guest. Protected copyright. nine). The very high exposures in studies 5 C and 6 (up to 211 mg/m3) also represented 25 selected operators who had incidental peak oo exposures due to accidental spillage during maintenance operations. In 68% of the 20 samples collected in study 5 phenol measure- .50 Q were also carried out. The correlation -Ea) ments Cu-a) of phenol with tt-MA as well as with S-PMA CuC._ 15 a) was rather poor: there were several increased a) 1-00 ._ values of phenol ( ) 50 mg/g creatinine) with- C) out concurrent exposure or substantially ._o increased tt-MA or S-PMA, and occasionally - 10 phenol was not increased (s< 5 mg/g creati- 02 nine) although both tt-MA and S-PMA were .50 CD clearly increased (tt-MA ) 1 mg/g creatinine E or S-PMA > 45 ,ug/g creatinine). During operations in oil refineries with aromatic plants (studies 9 and 10) exposures were gen- erally well below 1 ppm. In all cases distinct increases in both tt-MA and S-PMA were Time (h) found in the samples collected at the end of with those collected at the Figure 1 Urinay excretion oftn-MA and S-PMA in two workers (A and B) after the shift compared exposure to benzene on eight consecutive days. beginning. 616 Boogaard, Sittert

URINARY EXCRETION OF tt-MA AND S-PMA the dose that was excreted as tt-MA was 3 9% OVER TIME (range 1-9%-7-3%) and as S-PMA was 0-11% Occup Environ Med: first published as 10.1136/oem.52.9.611 on 1 September 1995. Downloaded from Figure 1 depicts the urinary excretion v time (range 0 05%-0 26%). plots from the two workers from study 8, who were biologically monitored for eight consecu- RELATION BETWEEN DIFFERENT URINARY tive days. The results of the personal air mea- METABOLITES OF BENZENE surements for benzene are shown in the In fig 2 the tt-MA concentrations were plotted graphs as bars. The first worker (A) was against the S-PMA concentration in the same exposed to relatively high concentrations of urine sample, both after logarithmic transfor- benzene up to 5-2 mg/M3 (1-6 ppm) on the mation, for all 188 workers from studies 1-12. third working day. The second worker was If for a worker more than one set ofvalues was exposed to much lower concentrations of ben- measured, the set of values corresponding to zene: the highest exposure was recorded on the highest airborne benzene concentration the sixth working day and was 1-63 mg/m3 was chosen, or if no air measurements were (05 ppm). It is clear, especially from the performed, the set with the highest values. A excretion profile of the more highly exposed highly significant correlation between tt-MA worker, that the excretion of tt-MA and S- and S-PMA was found (r = 0 795, P < PMA follow roughly the same pattern and 0-0001). Nevertheless, some deviations were that the excretion of both metabolites reflect found. In some cases where no exposure to the exposure to benzene. It is also clear that, benzene had occurred, relatively high concen- due to the shorter apparent half life of tt-MA trations of tt-MA (up to 0 71 mg/g creatinine) than S-PMA, tt-MA returns to baseline values were measured in the urine (without high at the beginning of the next shift after expo- concentrations of S-PMA: < 1 yg/g creati- sures up to 1 ppm whereas S-PMA tends to nine). accumulate at exposures over about 05 ppm. An exceptional finding was made in study 12. The two workers involved in this study TOXICOKINETICS OF tt-MA AND S-PMA carried out the same job in close vicinity of Twenty nine workers provided urine samples each other. Figure 3 shows the results of the on at least two consecutive days. For 27 work- personal air monitoring and the values of ers values for the apparent urinary excretion S-PMA and tt-MA. It seems that on the first constants could be calculated from 52 and 38 day the airborne concentrations of benzene data points for tt-MA and S-PMA, respec- are more or less identical for both men. One tively. The median k, for tt-MA was 0-136 1/h of the workers (closed symbols) excreted both and for S-PMA 0077 1/h. These values corre- S-PMA and tt-MA in concentrations that spond to median apparent elimination half match the concentration of benzene in air. In lives of 5-1 (mean (SD) 5 0 (2 3)) hours and the urine of the other man tt-MA was found 9 0 (9-1 (3 7)) hours, respectively. For 14 as expected, but surprisingly no increases in workers sufficient personal air monitoring and S-PMA were found. From the medical biological monitoring data were collected to records it seemed that this worker was treated allow an estimation of the urinary excretion of for epilepsy with high doses of carbamazepine, tt-MA and S-PMA as a percentage of the dose a known inducer of hepatic mixed function http://oem.bmj.com/ of inhaled benzene. The average percentage of oxidases.

Figure 2 Relation Regression line; r= 0.795: P<0.0001

between urinary given by: log (tt-MA (mg/g creatinine)) = on September 27, 2021 by guest. Protected copyright. concentrations of tt-MA 0.836 x log (S-PMA ('g/g creatinine)) and S-PMA in 188 -1-170 * tt-MA worker A workers exposed to -- 95% Cls (individual) L tt-MA worker B benzene. * S-PMA worker A --- 95% Cls (group) o S-PMA worker B O b, -nAlAA__A .,_1 A - on IvvA benzene worker A ^^ 2 -- 8 h TWA benzene worker B E

25 - 1 60 E

z C .C C: C~~~~~~~~~~~20~~~~~~~~~~~~~. .4_ C'1 - , 1.20 c 15- 0)

-0) C.)~~~~~~~~~~ E =. 10 --, cn

0 -j 0 8 16 24 32 40 48 560.00 < Time (h) -2 o 1 2 3 Figure 3 Urinary excretion of tt-MA and S-PMA in two workers (A and B) after exposure to benzene on two Log S-PMA (,ug/g creatinine) consecutive days. Biological monitoring ofexposure to benzene: a comparison between S-phenylmercapturic acid, trans,trans-muconic acid, and phenol 617

Figure 4 Relation - Regression line between respiratory 8 h -- Occup Environ Med: first published as 10.1136/oem.52.9.611 on 1 September 1995. Downloaded from TWA exposure to benzene 95% Cis (individual) and concentrations of tt- --- 95% Cis (group) MA and S-PMA in urinary samples collected at the end ofan 8 h shiftfrom 58 workersfrom studies 5, A 6, 10, and 12.

{D

m 0) N C 0)

200 400 600 800 1000 1200 -0 5 10 15 20 25 30 35 40 S-PMA (,ug/g creatinine) tt-MA (mg/g creatinine)

CORRELATION OF AIRBORNE BENZENE AND benzene in air (mg/m,3 8 h TWA) = 0.0758 x URINARY EXCRETION OF METABOLITES urinary S-PMA (ug/g creatinine) - 0-317, The relation between airborne benzene expo- r = 0-968 sures (8 h TWA) and urinary S-PMA and benzene in air (mg/m,3 8 h TWA) = 2.38 x was tt-MA concentrations examined in urinary tt-MA (mg/g creatinine) - 0 900, samples collected at the end of the shift from r = 0 959 58 workers in studies 5, 6, 10, and 12. The actual working hours in these studies were From the regression lines it was calculated close to eight hours. Data from workers who that after an 8 h TWA benzene exposure of 1 had worn respiratory protection during the ppm (3-25 mg/M3), the average concentration monitoring period were excluded from the in a urine sample collected at the end of the statistics. In cases where more than one data shift would be 47 ug S-PMA/g creatinine and point had been generated for a worker, the 1-74 mg tt-MA/g creatinine. The correspond- data point with the highest airborne benzene ing 95% confidence intervals (95% CIs) for concentration was used for the linear regres- the group means are 34-61 ,ug S-PMA/g crea- sion analysis. Highly significant correlations tinine and 1-2-2-2 mg tt-MA/g creatinine, http://oem.bmj.com/ (P < 0-0001) were found for both S-PMA respectively. and tt-MA (fig 4). The equations for the In the other studies where air measure- regression lines were: ments were made (1-4 and 8) the workers on September 27, 2021 by guest. Protected copyright. Figure 5 Relation Regression line between respiratory 12 h TWA exposure to benzene -- 95% Cls (individual) and concentrations of tt- --- 95% Cls (group) MA and S-PMA in urinary samples collected at 20 the end of a 12 h shiftfrom ,I / 28 workersfrom studies / 1-4 and 8.

15

V) E

E 10

N a) c

An - v I 0 50 100 150 200 250 300 350 400 450 500 0 2 4 6 8 10 12 S-PMA (,ug/g creatinine) Zt-MA (mg/g creatinine) 618 Boogaard, Simert

were on duty in 12 hour shifts. A separate established biological monitoring method regression analysis was made with the data with the more recent biological monitoring Occup Environ Med: first published as 10.1136/oem.52.9.611 on 1 September 1995. Downloaded from collected in these studies. Again, only one set methods. of data was included for each worker and data We found that measurement of tt-MA as collected while respiratory protection was described by Ducos and coworkers sometimes worn were excluded. Also, data collected dur- gave non-reproducible results.56 It turned out ing operations with apparent dermal expo- that this phenomenon was due to differences sure, as recorded by the occupational in the quaternary ammonium ion exchange hygienist, were excluded from the analyses. columns. With some brands the recovery of tt- Finally, the following regression equations MA from the columns seemed to be depen- (fig 5), with highly significant correlations dent on the pH of the urine sample. (P < 0-0001), were obtained by analysing Alkalisation of the samples to a pH of 7 to 10 data sets from 28 workers: gave optimum recoveries that were invariably over 95%. It seemed that the solid phase benzene in air (mg/m,3 12 h TWA) = 0-0507 extraction was more efficient at a higher pH, x urinary S-PMA (ug/g creatinine) + 0-518, which resulted in cleaner extracts and a lower r = 0 959 limit of detection (0-01 mg/l). Recently, this same favourable effect of alkalisation was also benzene in air (mg/M,3 12 h TWA) = 1.98 x reported by Lee and coworkers who added 2 urinary tt-MA (mg/g creatinine) + 0 390, volumes of a TRIS buffer (pH 10) to the r = 0-862 urine samples before the solid phase extrac- tion.9 Overall, the method proved to be simple, straightforward, and robust, which Discussion resulted in small variation coefficients within Because of the potential adverse health and between runs. effects, including acute myeloid leukaemia, In the method used for the determination that have been associated with long term of S-PMA, the internal standard S-BMA was occupational exposure to high concentrations replaced by deuterated S-PMA in the present of benzene (100-500 ppm), it is important to study for two reasons. The first reason was to minimise exposure to benzene. To evaluate improve the sensitivity and reliability of the the effectiveness of control measures, biological method. In our previous report it was stated monitoring techniques are needed that allow that the limit of detection varied from 1 to detection of benzene to well below 1 ppm. 5 jug/l, depending mainly on matrix effects.'3 It Due to the high and variable background of seemed that in urine specimens with a high urinary phenol caused by endogenous pro- specific gravity, S-PMA concentrations lower duction of phenol as well as high dietary and than about 5 ,ug/l could in certain cases not be environmental exposure to phenol, urinary attributed with certainty. With deuterated S- phenol is not a suitable variable for biological PMA as the internal standard this problem monitoring of exposure to benzene of below was solved and a limit of detection of 1 ,ug/l 5 ppm. In the past few years, tt-MA and S- could be guaranteed. PMA, two minor metabolites of benzene, The second reason was that toluene was http://oem.bmj.com/ have been proposed to replace phenol as bio- recently reported to be metabolised in markers of exposure to benzene. These humans to S-BMA. 19 20 In many industrial metabolites can both be measured in urine, processes with a risk of exposure to benzene which in our view is a prerequisite for good co-exposure to toluene may occur and bio- biological monitoring. Since its introduction transformation of toluene to S-BMA might in 1985, improved methods have been pub- interfere with its use as an internal standard. lished for the measurement of urinary tt-MA In some preliminary studies we could indeed by Inoue and co-workers4 who claimed to be identify small amounts of S-BMA (up to 4 0 on September 27, 2021 by guest. Protected copyright. able to detect benzene exposure to 6-7 ppm ,ug/l) in 10 workers exposed to low concentra- and by Ducos and co-workers5 6 who tions of toluene (up to 5-3 ppm, 8 h TWA) improved Inoue's method by increasing the (unpublished observations). We found in this sensitivity, which allowed biological monitor- as well as in our previous studies'3 that only ing of benzene exposure down to 1 ppm (8 h 0-11% of inhaled benzene is metabolised to TWA). Recently, we published an improve- S-PMA and although toluene is probably ment of the method for the determination of metabolised even less to S-BMA, there might S-PMA as proposed by Stommel and co- still be serious interference in some cases as workers'2 that allows routine biological moni- the airborne concentrations of toluene may toring of exposure to benzene as low as 0 3 exceed those of benzene by one or two orders ppm (8 h TWA).'3 of magnitude in many industrial operations. This paper reports the results of 12 sepa- The problem of interference due to co-expo- rate investigations where measurements of sure to toluene may also be overcome with a both urinary tt-MA and S-PMA were applied. halogenated phenylmercapturic acid, such as The methods were compared for sensitivity, S-4-fluorophenylmercapturic acid, as the specificity, and ease of operation, to get a bet- internal standard,'2 1" but this will not solve ter insight into their suitability for routine bio- the loss of sensitivity due to matrix effects. logical monitoring of industrial exposure to Because of the increased sensitivity of the low concentrations of benzene. To a limited measurement of S-PMA, the benzene uptake extent urinary phenol measurements were due to cigarette smoking was detectable in all also performed to enable a comparison of this 14 smokers from the control group (mean Biological monitoring ofexposure to benzene: a comparison between S-phenylmercapturic acid, transtrans-muconic acid, and phenol 619

(SEM) 3-61 (0-57) ,ug/g creatinine). In 20 out sion lines indicate that the reliability of tt-MA of the 38 non-smokers of the control group and S-PMA as biomarkers for benzene expo- Occup Environ Med: first published as 10.1136/oem.52.9.611 on 1 September 1995. Downloaded from S-PMA could also be detected albeit in sure during eight hour shifts are almost identi- much lower concentrations (1 99 (0 32) ,ug/g cal. It is obvious, however, that the reliability of creatinine). The sensitivity of the tt-MA tt-MA as a biomarker for benzene exposure measurement was too low to allow detection during 12 hour shifts is much less than for of tt-MA in most non-smokers and some S-PMA, which is totally due to its shorter smokers in the control group. The concentra- elimination half life. tions of tt-MA in the smokers (0-058 (0-013) From these correlations it follows that mg/g creatinine) and non-smokers (0 037 exposure to 1 ppm benzene (8 h TWA) will (0-16) mg/g creatinine) were relatively low lead to an average excretion of 1 74 mg tt- compared with the results of a recent study on MA/g creatinine and 47 ug S-PMA/g creati- tt-MA conducted in Singapore where values nine in urine samples collected at the end of of 0-19 and 0-14 mg/g creatinine were found the shift. These values seem to be confirmed in 35 smokers and 23 non-smokers, respec- by the regression equation in fig 2, from tively.9 Although the same detection limit for which it can be calculated that 47 pg S- tt-MA was reported as in our study, tt-MA PMA/g creatinine corresponds to 1 69 mg tt- was detectable in all smokers and almost half MA/g creatinine. The value of 47 Pg/g of the non-smokers. The investigators thought creatinine for S-PMA also agrees well with the that their relatively high background values findings from our previous studies, where a were due to environmental pollution. In similar value was found.'3 Bechtold and co- another recent study values of 0-29 and 0 09 workers found in a group of 14 Chinese mg tt-MA/g creatinine were measured in 42 female workers that exposure to 4-4 ppm ben- smokers and 42 non-smokers, respectively. zene (8 h TWA) led to an average excretion of Although the smokers in these two studies 6-2 mg tt-MA/g creatinine in urine samples seemed to consume larger quantities of ciga- collected at the end of the shift. This matches rettes than the moderate smokers in our con- our results extremely well as according to the trol group, there was no correlation between regression line obtained with our data such an the number of cigarettes smoked and excre- exposure would lead to excretion of 6-4 mg/g tion of tt-MA,9 but a weak correlation was creatinine. Our findings are also in good found between excretion of tt-MA and urinary agreement with two other studies in 23 cotinine concentrations. French and 38 Belgian male workers where it The toxicokinetics, based on spot samples, was found that exposure to 1 ppm benzene were studied for both tt-MA and S-PMA. The (8 h TWA) corresponded to 1 mg tt-MA/16 mean (SD) apparent urinary elimination half and 1-4 mg n-MA/g creatinine,'0 respectively. life (ti,,) for S-PMA was 9 1 (3.7) hours, In some cases skin contact with liquids con- which is almost identical to the t,12 of 9 0 (4 5) taining benzene was reported by the industrial hours previously reported.'3 No discrete val- hygienist. The personal air monitoring data ues for the tl,2 of tt-MA have been reported as and the concentrations of tt-MA and S-PMA yet, but it has been found that in humans tt- measured in the urine collected at the end of MA showed a t1,2 similar to that of phenol.5 the shift strongly suggest that in these cases http://oem.bmj.com/ Our data corroborate this finding: the esti- rather extensive dermal uptake of benzene mated tl/2 for tt-MA was 5 0 (2-3) hours. The had taken place. For example, during the average percentages of the doses of inhaled replacement with bare hands of a plunger in a benzene that were excreted in the urine as tt- skimmerpump for natural gas condensate in MA and S-PMA were 3 9% (range study 1, an airborne benzene concentration of 1-9%-7-3%) and 0-11% (range 0 05%- 19-2 mg/M3 was measured as a 12 h TWA. In the urine collected at the end of the shift 12-5 0 26%), respectively. Our value for the excre- on September 27, 2021 by guest. Protected copyright. tion of tt-MA is about twice the only value of mg tt-MA and 798,ug S-PMA/g creatinine 1-9% reported so far.4 This difference is partly were measured. According to the regression explained by the lower respiration rate of 09 equations for the relation between airborne m3/h used for the estimation by Inoue and benzene (12 h TWA) and the urinary meta- coworkers. The value found for S-PMA con- bolites in urine at the end of the shift, the con- firms our previous findings of a mean excre- centration tt-MA and S-PMA would indicate tion of 0 1 1% as S-PMA with a range of 0 05 exposure to 25-2 and 41-0 mg benzene/ml, to 0 29%.'3 Most probably these rather wide respectively. Although the value for tt-MA is ranges for both tt-MA and S-PMA are partly just inside the 95% CI for individual results, it due to the fact that the calculations are based is extremely unlikely (P < 0-001) that the on spot samples, with the inherent risk of result for S-PMA is solely due to inhalation under and overestimation of areas under the exposure. curve. The present study was undertaken to com- Despite the fact that only spot samples pare the suitabilities of tt-MA and S-PMA for were collected and that tl/2 of tt-MA is consid- biological monitoring of low level benzene erably shorter than that of S-PMA a good exposure. Our results confirm previous find- correlation was found between urinary con- ings by ourselves and others that the measure- centrations of tt-MA and S-PMA (fig 2). Also ments of both urinary tt-MA and S-PMA are good correlations between benzene in air and sensitive tests for monitoring exposure to low S-PMA as well as tt-MA were found (figs 4 concentrations of benzene. We confirmed the and 5). The 95% CIs in these figures as well suitability of S-PMA to detect exposures to as the correlation coefficients for the regres- benzene as low as 0 3 ppm (8 h TWA). This 620 Boogaard, Sittert

low level is fully due to the high specificity of lation. Int Arch Occup Environ Health 1992;64:179-84. 4 Inoue 0, Seiji K, Nakatsuka H, Watanabe T, Yin SN, Li Occup Environ Med: first published as 10.1136/oem.52.9.611 on 1 September 1995. Downloaded from the S-PMA test. The increased sensitivity GL, et al. Urinary tt-muconic acid as an indicator of achieved by lowering the detection limit led to exposure to benzene. BrJ'Ind Med 1989;46:122-7. 5 Ducos P, Gaudin R, Robert A, Francin JM, Maire C. possible detection of exposure to benzene Improvement in HPLC analysis of urinary transtrans- caused by moderate smoking. Despite the fact muconic acid, a promising substitute for phenol in the assessment of benzene exposure. Int Arch Occup Environ that we were able to improve the method of Health 1990;62:529-34. measurement of urinary tt-MA by a more effi- 6 Ducos P, Gaudin R, Bel J, Maire C, Francin JM, Robert A, et al. transtrans-Muconic acid, a reliable biological cient clean up, we also confirmed the conclu- indicator for the detection of individual benzene expo- sions by others that with tt-MA, at best, sure down to the ppm level. Int Arch Occup Environ Health 1992;64:309-13. exposures of over 05 ppm can reliably be 7 Bechtold WE, Lucier G, Birnbaum LS, Yin SN, Li GL, detected.'014 The reason for this is that in Henderson RF. Muconic acid determinations in urine as a biological exposure index for workers occupationally unexposed people relatively high background exposed to benzene. Am Ind HygAssocJ 1991;52:473-8. concentrations of tt-MA are consistently 8 Johnson ES, Lucier G. Perspectives on risk assessment impact of recent reports on benzene. Am _J Ind Med found.8 0112' This background may possibly 1992;21:749-57. be due to consumption of sorbic acid, which 9 Lee BL, New AL, Kok PW, Ong HY, Shi CY, Ong CN. Urinary transtrans-muconic acid determined by liquid is used as a preservative in many foodstuffs.5 chromatography: application in biological monitoring of There may be a preference for uA-MA over benzene exposure. Clin Chem 1993;39:1788-92. 10 Lauwerys RR, Buchet JP, Andrien F. Muconic acid in S-PMA as a biomarker for benzene exposures urine: a reliable indicator of occupational exposure to of more than 1 ppm (8 h TWA) as HPLC is benzene. Am I Ind Med 1994;25:297-300. 11 Rauscher D, Lehnert G, Angerer J. Biomonitoring of easier to operate than gas chromatography- occupational and environmental exposures to benzene mass spectroscopy and in most cases is also by measuring transtrans-muconic acid in urine. Clin Chem 1994;40:1468-70. more readily available. Due to the rather 12 Stommel P, Muller G, Stucker W., Verkoyen C, Schobel S, expensive solid phase extraction, the cost/ Norpoth K. Determination of S-phenylmercapturic acid in the urine-an improvement on the biological moni- sample of analysis of tt-MA is about equal to toring of benzene exposure. Carcinogenesis 1989;10: that of S-PMA. A disadvantage of tt-MA for 279-82. 13 Van Sittert NJ, Boogaard PJ, Beulink GDJ. Application of biological monitoring of exposure to benzene the urinary S-phenyimercapturic acid test as a biomarker is that its formation is decreased in the event for low levels of exposure to benzene in industry. Br 7 Ind Med 1993;50:460-9. of co-exposure to other aromatic hydro- 14 Popp W, Rauscher D, Muller G, Angerer J, Norpoth K. carbons,22 which may lead to underestimation Concentrations of benzene in blood and S-phenylmer- capturic acid and t,t-muconic acid in urine in car of benzene exposure. Although we have not mechanics. Int Arch Occup Environ Health 1994;66:1-6. yet been able to confirm it, our findings in 15 CONCAWE's petroleum products and health manage- ment groups. Gasolines. Product dossier 921103. Brussels: study 12, where hardly any S-PMA was found CONCAWE, 1992. in a man treated with high doses of an inducer 16 Rumpf P. Darstellung, Charakteriserung und Trennung der Merkaptursauren von Benzol, Toluol, Ethylbenzol of the hepatic mixed function oxidase, might und Xylolen durch GC-MS Analyse. Hamburg: indicate that the formation of S-PMA may be Universitgt Hamburg, 1981. (PhD Thesis.) 17 NIOSH. Back up data report benzene, method S311. Manual influenced by co-exposure as well. of analytical methods, 2nd ed. Vol 3. Cincinnati: DHEW (NIOSH), 1977. (Publication No 77-157C.) The technical assistance by H van der Waal, J Kweekel, and 18 Doctor JH, Zielhuis R. Phenol excretion as a measure of Y N Vreugd (Shell Biomedical Laboratory; Shell Nederland occupational exposure. Ann Occup Hyg 1967;1O:317-36. Raffinaderij/Chemie B V, Rotterdam, The Netherlands) is 19 Takahashi S, Matsubara K, Hasegawa M, Akane A, gratefully acknowledged. Shiono H. Detection and measurement of S-benzyl-N- acetylcysteine in urine of toluene sniffers using capillary gas chromatography. Arch Toxicol 1993;67:647-50. http://oem.bmj.com/ 20 Takahashi S, Kagawa M, Shiwaku K, Matsubara K. 1 Yardley-Jones A, Anderson D, Parke DV. The toxicity of Determination of S-benzyl-N-acetyl-L-cysteine by gas benzene and its metabolism and molecular pathology in chromatography/mass spectrometry as a new marker of human risk assessment. BrJInd Med 1991;48:437-44. toluene exposure. JAnal Toxicol 1994;18:78-80. 2 Angerer J, Scherer G, Schaller KH, Muller J. The determi- 21 Melikian AA, Prahalad AK, Hoffmann D. Urinary nation of benzene in human blood as an indicator of trans,trans-muconic acid as an indicator of exposure environmental exposure to volatile aromatic com- to benzene in cigarette smokers. Cancer Epidemiol pounds. Fresenius Journal of Analytical Chemistry 1991; Biomarkers Prev 1993;2:47-51. 339:740-2. 22 Inoue 0, Seiji K, Watanabe T, Kasahara M, Nakatsuka H, 3 Brugnone F, Perbellini L, Maranelli G, Romeo L, Yin S, et al. Mutual metabolic suppression between ben-

Guglielmi G, Lombardini F. Reference values for blood zene and toluene in man. Int Arch Occup Environ Health on September 27, 2021 by guest. Protected copyright. benzene in the occupationally unexposed general popu- 1988;60:15-20.